On Tuesday, November 22, 2022 2:32:11 PM EST Sandro-Alessio Gierens wrote: > Package: wnpp > Severity: wishlist > Owner: Sandro-Alessio Gierens <sandro@gierens.de> > > * Package name : ranges > Version : 1.0.0 > Upstream Author : Sandro-Alessio Gierens <sandro@gierens.de> > * URL : https://github.com/gierens/ranges > * License : GPLv3 > Programming Lang: C > Description : Command line program to extract ranges from various > types of lists, e.g. integer numbers, dates, IP and MAC addresses. > > ranges is a command line program written in C that extracts ranges from > various types of lists. By default it parses signed decimal integer > lists, but given the right argument it can work with unsigned > hexadecimal, octal and binary numbers, dates, IPv4, IPv6 and MAC > addresses. The list input is given over the standard input, so by pipe, > and is assumed to be sorted, but can have duplicates. > > Relevance > > I work in a data center and recently had the problem that I needed to > find out which IP addresses of a subnet were not yet assigned in > a /etc/hosts file. Because there were already too many addresses to > get a good overview, I began to wonder if there was any command line > tool that would allow me to compile the list of IPs into a list of > IP ranges, so the gaps and their size would be obvious. Unfortunately > I only found stack overflow discussions suggesting writing a script, > and this is what I did to back then too. While this usually doesn't > take more than a couple of minutes, ranges has too advantages: > It already implements a bunch of different list types including > nasty things like dates for example, and therefore would spare > people from replementing such scripts over and over again. Aside > from that it is written in C and therefore fast. According to my > tests it is, depending on the machine, 20 to 40 times faster than > a comparable Python3 script. It can crunch 130 MB of IPv4 addresses > in a second. > > Eventhough I just published the initial release I've been working on > this for a couple of weeks now and extensively checked that it is > stable and secure. My test suite consists of 185 tests that verify > the correct functionality of each mode and argument. Each test is > first run without and then with valgrind memcheck, so there should > not be any leaks or other memory errors. > > Maintainance > > As the upstream author of the software I would also maintain the > package. This would be my first package, but I already have a make > rules to build a deb package and check it with lintian, so I'm > not unprepared :) The package name is very generic. I would pick something more specific. Also, I suspect the speed advantage would be less significant if you used SubnetTree (python3-subnettree), which does all the hard work in C. Is this really needed? Scott K
Attachment:
signature.asc
Description: This is a digitally signed message part.